DISCRIMINATION AND COMPARISON. 543 



have given measurements by method 1 which cover the 

 most extensive series, and moreover apply to six different 

 colors of light. These experiments (performed in Helm- 

 holtz's laboratory, apparently,) ran from an intensity called 

 1 to one which was 100,000 times as great. From intensity 

 2000 to 20,000 Weber's law held good ; below and above 

 this range discriminative sensibility declined. The incre- 

 ment discriminated here was the same for all colors of 

 light, and lay (according to the tables) between 1 and 2 per 

 cent of the stimulus.* Delboeuf had verified Weber's law 

 for a certain range of luminous intensities by method 4 ; 

 that is, he had found that the objective intensity of a light 

 which appeared midway between two others was really the 

 geometrical mean of the latter's intensities. But A. Lehmann 

 and afterwards Neiglick, in Wundt's laborator}^, found that 

 effects of contrast j^layed so large a part in experiments 

 performed in this way that Delboeuf's results could not be 

 held conclusive. Merkel, repeating the experiments still 

 later, found that the objective intensity of the light which 

 we judge to stand midway between two others neither 

 stands midway nor is a geometric mean. The discrepancy 

 from both figures is enormous, but is least large from the 

 midway figure or arithmetical mean of the two extreme in- 

 tensities, f Finally, the stars have from time immemorial 

 been arranged in ' magnitudes ' supposed to differ by equal- 

 seeming intervals. Lately their intensities have been 

 gauged photometrically, and the comparison of the subjec- 

 tive with the objective series has been made. Prof. J. Jas- 

 trow is the latest worker in this field. He finds, taking 

 Pickering's Harvard photometric tables as a basis, that the 

 ratio of the average intensity of each ' magnitude ' to that 

 below it decreases as we pass from lower to higher magni- 

 tudes, showing a uniform departure from Weber's law, if 

 the method of equal-appearing intervals be held to have 

 any direct relevance to the latter.:}: 



* Berlin Acad. Sitzungsberichte, 1888, p. 917. Other observers (Dobro. 

 wolsky, Lamansky) found great differences in different colors. 



f See Merkel's tables, loc. cit. p. 568. 



X American Journal of Psychology, i. 125. The rate of decrease is 

 small but steady, and I cannot well understand what Professor J. means by 

 saying that his figures verify Weber's law. 



